Probing Membrane Degradation with Varying Catalyst Layer Thickness and Carbon Support

Abstract

Understanding membrane degradation continues to be of significant importance when studying fuel cell durability. Fuel cell performance and durability depends heavily on the membrane’s chemical and mechanical integrity. In Nafion® membranes, hydroxyl radicals are the primary cause of chemical membrane degradation as they attack the membrane’s polymer structure. Radical formation can ensue via peroxide decomposition at the cathode or gases reacting on Pt particles within the membrane.[1] A previous research effort on membrane degradation as a function of ultra-low Pt loadings (≤ 0.1 mgPt cm-2) showed the degradation rate increases with Pt loading.[2] This is most likely due to more heterogenous sites for radical formation. Manufacturing MEAs with ultra-low Pt loadings can yield thin, non-uniform catalyst layers. While thinner catalyst layers typically yield higher Pt utilization, they can also induce flooding issues because of its lower pore capacity. Conversely, thicker electrodes can provide a more uniform catalyst layers, however the Pt activity can be adversely impacted due to poorer gas permeability within thicker catalyst layer.[3]Characteristics of the catalyst layer structure such as catalyst layer thickness, ionomer to carbon (I/C) ratio, and carbon support type can influence fuel cell cost, performance, and durability. An optimal I/C ratio is critical to achieve better overall fuel cell performance. Low I/C ratio can inhibit proton transport while higher ratios tend to block catalyst sites, thus hindering gas diffusion. The carbon support’s porous structure influences catalyst utilization and electrochemical kinetics. In this research effort we studied membrane degradation with ultra-low Pt loadings while varying catalyst layer thickness, I/C ratio, and types of carbon support.Twenty-five cm2 membrane electrode assemblies (MEAs) were prepared by spraying catalyst inks on to Nafion® membranes (NR211, Ion Power, Inc.) using a Sono-Tek Exactacoat spray coater. The catalyst inks consisted of TKK catalyst mixed with 5% ionomer solution (Dupont D521). Catalyst layer thickness was modified by using 20% and 50% Pt/C catalyst. I/C ratios were chosen to maintain the ionomer to Pt and ionomer to carbon ratios the same with respect to the 50% Pt/C catalyst. I/C ratios of a 0.22 and 0.8 were used in a 20% Pt/C catalyst. Two types of carbon supports were applied throughout this study: a porous support, Ketjen black and a solid support, Vulcan XC-72.We applied a DOE membrane chemical degradation accelerated stress test (AST) to promote the generation of hydroxyl radicals.[4] The AST test conditions require the fuel cell to be held at open circuit voltage while at 90oC, 30% relative humidity and 150 kPa backpressure. Hydrogen (H2) and air flows were set at 10 stoics at 0.2 A/cm2. A voltage recovery protocol (RP) was also employed every 24 hours to decouple irreversible voltage losses from reversible voltage losses.Fuel cell performance was evaluated every 24 hours to probe membrane degradation. Electrochemical techniques such as linear sweep voltammetry was employed to assess H2 permeation through the membrane. In addition, we conducted cyclic voltammetry to calculate electrochemical active surface area (ECSA). Fuel cell effluent water was collected for further scrutiny of membrane degradation by probing for fluoride and sulfate ions using Ion Chromatography (IC). Fluoride emission rates (FER) and sulfate emission rates (SER) were used to gauge the condition of the MEA during its lifetime.Figure 1A shows H2 crossover measurements for three samples. H2 permeation increased with time for both MEA samples with I/C = 0.8, MEA 50% and MEA 20% 0.8. The MEA using the 50% catalyst exhibited a higher crossover current density at End of Test (EOT) than MEA 20% 0.8 even though the latter was tested for a longer time. MEA 20% 0.2 showed a crossover rate that declined over time. This unusual phenomena could be related to a delaminated electrode, hence current cannot be measured using this technique. IC results for FER are shown in Figure 1B. MEA 50% had an increasing FER throughout testing, while MEA 20% 0.8 produced a lower FER under AST. For MEA 20% 0.2 there was a minimal amount of fluoride emitted over time indicating a lack of reaction sites to generate radicals. ECSA gradually declines with time with only partial surface area restoration after the RP was applied. This talk will summarize the results of membrane degradation with varying catalyst layer thicknesses, ionomer concentrations, and types of carbon supports.[1] M.P. Rodgers, L.J. Bonville, H.R. Kunz, D.K. Slattery, J.M. Fenton, Chem. Rev. (2012).https://doi.org/10.1021/cr200424d.[2] A. Spears, T. Rockward, R. Mukundan, F.H. Garzon, ECS Trans. . 92 (2019) 467–475.https://doi.org/10.1149/09208.0467ecst.[3] M.S. Wilson, S. Gottesfeld, J. Appl. Electrochem. (1992). https://doi.org/10.1007/BF01093004.[4] https://doi.org/http://www1.eere.energy.gov/hydrogenandfuelcells/fuelcells/pdfs/component_durability_profile.pdf Figure 1